Large-scale system-level digitalisation initiatives in the National Health Service in England: insights from three national evaluations

Why this matters for everyday healthcare

Across England, the National Health Service (NHS) has poured billions of pounds into computers, networks, and artificial intelligence with the promise of safer, more joined-up care. Yet many patients and staff still experience lost letters, repeated tests, and systems that do not talk to each other. This article looks back over 15 years of major national digital projects and asks a simple question with big implications: what actually worked, what did not, and how can future efforts finally deliver the smooth, data-driven care people have been promised?

The rise of big digital projects

To answer this, the authors combined three large, independent evaluations of national NHS programmes worth around £13 billion in total. The first was an early attempt to roll out shared electronic records to all hospitals. The second funded “exemplar” hospitals to become digital trailblazers and spread good practice. The third backed artificial intelligence tools and the rules and support needed to use them safely. Together, these evaluations drew on over a thousand interviews, hundreds of hours of observing clinics and meetings, and more than two thousand internal documents collected between 2009 and 2024. This unusually long and detailed view allowed the researchers to see patterns that would be invisible in shorter, single-project studies.

Hidden foundations and stubborn plumbing

A clear message was that basic digital “plumbing” is both essential and often neglected. Reliable Wi‑Fi, dependable electronic records, and clean, well-structured data turned out to be the quiet workhorses on which everything else depends. Where networks were patchy, systems unreliable, or data poorly organised, newer tools—including AI—struggled or failed outright. At the same time, trying to link many old and new systems across hospitals, GP practices, and social care created huge headaches. Different suppliers did not always cooperate, standards were unevenly applied, and differences in how care was delivered meant that even when data could move, it was not always useful or safe. These long‑standing infrastructure and interoperability problems repeatedly slowed progress.

People, politics, and shifting goals

Technology itself was rarely the only obstacle. More often, difficulties grew from how people, organisations, and national bodies interacted around that technology. Ministers announced bold deadlines and sweeping visions, raising expectations among staff that change would be rapid and pain‑free. In reality, large system changes in busy hospitals and clinics take years, not months. Programmes were launched quickly, sometimes without solid baseline measurements or realistic plans, and then were reshaped or cut short as political priorities, senior leaders, and central agencies changed. This unstable governance meant that objectives drifted, support structures were reorganised, and hard‑won lessons were easily lost, even as the same problems resurfaced in the next initiative.

Learning that travels too slowly

When digital projects did work well, it was often because of strong local leadership, careful engagement with frontline staff, and chances to learn from peers facing similar challenges. The “exemplar” hospitals, for example, were able to adapt systems to their context and then share blueprints and informal know‑how with others, speeding up adoption elsewhere. However, the study found that such learning was rarely captured and reused in a systematic way. National evaluations sometimes generated rich insights but these were not consistently fed back into day‑to‑day decision making, and there was little continuity of learning from one major programme to the next. As a result, new initiatives tended to repeat old mistakes rather than build steadily on past experience.

A staged path to truly digital care

Drawing these strands together, the authors argue that national digital efforts should be seen not as one‑off projects but as an ongoing journey that must balance technology, people, organisations, and policy over decades. They propose a three‑stage approach they call the ILIAD model. First, invest in robust, shared infrastructure so that every part of the system has workable networks, records, and data. Second, create strong learning links between the most advanced sites and those catching up, so experience can spread quickly and be adapted locally. Third, once these foundations and learning networks exist, concentrate advanced innovation—such as AI tools—in settings that are ready to experiment safely and then scale what proves useful. For patients and staff, the promise of this approach is not flashy gadgets, but a steadier, more reliable path toward care where information follows the person, services are better coordinated, and new technologies genuinely make work easier and care safer.

Citation: Cresswell, K., Williams, R. Large-scale system-level digitalisation initiatives in the National Health Service in England: insights from three national evaluations. npj Digit. Med. 9, 301 (2026). https://doi.org/10.1038/s41746-026-02495-8

Keywords: digital health transformation, electronic health records, NHS England, healthcare interoperability, artificial intelligence in healthcare